Camera Labs

With Nikon introducing their FX-sensor there is a lot of expectation as to what you can expect from the D3 or other FF-sensor bodies like those from Canon in comparison to the APS-C sensor bodies.
With respect to photon-collection, everything seems clear: with 2x* the size of a single FF/FX-photo cell you collect double the photons in each cell.
This in turn means that with the same exposure you catch 2x the photons in one cell than with an APS-C size sensor cell and that is 2x the number of electrons that FF/FX-cell can generate. So s/n-ratio should be 2x better. All clear?
Not quite! Last time I checked, larger electronic structures generate inherently higher noise
But for the sake of simplicity and as all the results point to a 1 stop (i.e.2x) better s/n-ratio let's assume that is so.
This means that you get the same noise with a FF/FX-sensor at ISO 1600 than with an APS-C sensor at ISO 800. Or: You can expose -1EV on the FF/FX-sensor to get the same results as with the APS-C sensor.

Now on to dynamic range: With this -1EV exposure on the 2x larger photo-cell you should just capture the same amount of photons and generate the same amount of electrons from it as on the APS-C sensor with normal (0EV) exposure. But as the larger photosite has 2x the capacity for photons than the smaller cell it should be capable of having +1EV more dynamic range.

Has anyone any evidence confirming (or disprove) my deductions?
Any meaningful tests with Canon's FF-bodies on DR??
And what about larger electrical noise in larger electronic structures???

----------
*I'll stick to 2x in this article just for simplification! In reality the FF/FX_sensor has 2.5x the size of an APS-C sensor and each photosite is 2.5-3x larger!

_________________
Thomas (beware: Nikon-fanboy and moderator!) My Lens Reviews, My Pictures, My Photography Blog
D800+assorted lenses

Hi Thomas, I can reveal I've been doing some tests with a Stouffer T4110 wedge chart and have recorded images using a few cameras so far including the Canon 40D and 5D.

I realise a number of other sites also use the same chart with various software analysis, but I think there's value to even just posting histograms for comparison.

We're busy preparing several reviews right now - along with just posting our Sony A700 final update - so bear with us! But we'll have a look at these results in the near future.

In the meantime, anyone else please feel free to comment or contribute!

Gordon

I'm looking forward to your 40D/5D histogram comparison.


I wonder (but can't find out) whether increasing the area of a photosite also allows the three dimensional structure to grow a little? Is it possible that doubling the area means that they can also increase the cell depth allowing an even bigger full-well capacity than the doubling of your example? Provided read-out noise doesn't increase in proportion then dynamic range increases.

On another note, this Northlight Images page has the following speculation:

"Canon is working on a new generation CMOS sensor – lower voltage, full 100% cover micro-lenses and better dynamic range allowing 16 bit A/D per pixel raw files.

These will support a clean 6400 ISO similar to today's 1600 and go to 25600 with acceptable noise.

The next generation low voltage CMOS generally gives 1 to 2 stops lower noise at higher ISOs"

Unfortunately we apparently won't see this technology for a few years but it's nice to know that there will be alternatives to medium format cameras for those A1 prints!

Bob.

_________________
OM-D E-M5 + ED 12-50mm, H-PS14042E, H-F007014E, L-RS014150E, M.ZUIKO DIGITAL 45mm 1:1.8.
Gitzo GT1541T, Arca-Swiss Z1 DP ball-head.
Astrophotography: TEC 140 'scope, FLI ML16803 camera, ASA DDM60 Pro mount.